Super Micro, Vast Data, Kioxia and others are partnering to build purpose‑built AI infrastructure, aiming to meet the high‑performance demands of agentic AI workloads.

Four leading technology firms—Super Micro, Vast Data, Kioxia and a consortium of other innovators—have announced a joint effort to create “Neoclouds,” a purpose‑built AI infrastructure designed to handle the intensive compute and storage demands of emerging agentic AI workloads.

Why a New AI Infrastructure Is Needed

Agentic AI models, which can autonomously plan and execute tasks, require unprecedented levels of parallel processing, low‑latency networking, and high‑throughput storage. Existing cloud offerings often struggle to deliver the consistent performance needed for real‑time inference and large‑scale training, prompting industry players to explore specialized architectures.

Key Components of the Neoclouds Initiative

  • Super Micro will supply modular, high‑density server platforms optimized for AI accelerators.
  • Vast Data contributes its disaggregated storage fabric, offering petabyte‑scale capacity with sub‑millisecond latency.
  • Kioxia provides next‑generation NVMe SSDs designed for sustained high‑throughput workloads.
  • Partner ecosystem includes networking vendors delivering 400 Gbps Ethernet and custom interconnects.

The collaboration emphasizes open standards and composability, allowing enterprises to mix and match components based on specific workload requirements while maintaining a unified management layer.

Potential Impact on Enterprises

By delivering a tightly integrated stack, Neoclouds aims to reduce total cost of ownership for AI projects, shorten time‑to‑value, and enable new use cases such as autonomous decision‑making systems, real‑time language agents, and advanced robotics.

Early adopters are expected to benefit from improved scalability, as the architecture supports seamless expansion of compute nodes and storage tiers without disruptive re‑architecting.

“We’re building the foundation that will let the next generation of AI applications run at scale, reliably and efficiently,” said a spokesperson for Super Micro.

The Neoclouds project also aligns with broader industry trends toward “hyper‑converged” AI infrastructure, where compute, storage, and networking are tightly coupled to minimize bottlenecks.

SiliconANGLE coverage of Neoclouds AI infrastructure partnership